I primarily work on LLM fine-tuning and distillation, generating synthetic data, and creating Agents.https://djellalmohamedaniss.github.io/https://djellalmohamedaniss.github.io/posts/data-free-3bit-quantization/https://djellalmohamedaniss.github.io/posts/data-free-3bit-quantization/By trading heavy FP16 MatMuls for SRAM lookups and 1-bit additions, our custom quantization pipeline squeezes state-of-the-art models down to approx. 3 bits per weight with minimal accuracy loss. Here is how bypassing Tensor Cores could reshape the design of future edge AI chips.Thu, 16 Apr 2026 00:00:00 GMThttps://djellalmohamedaniss.github.io/posts/calibrating-llm-reasoning/https://djellalmohamedaniss.github.io/posts/calibrating-llm-reasoning/In this blog post, we will discuss about Controlling reasoning effort in LLMs ( the gpt-oss-style ) and Calibrating LLM Reasoning effort via Label-Free Alignment.Tue, 17 Feb 2026 00:00:00 GMThttps://djellalmohamedaniss.github.io/posts/stop-using-embeddings-for-everything-in-rag/https://djellalmohamedaniss.github.io/posts/stop-using-embeddings-for-everything-in-rag/Why deterministic query translation should often come before embeddings in enterprise RAG systems, and how to combine both in a hybrid approach.Thu, 11 Dec 2025 01:34:00 GMT